我們提出一個新的真假人臉的辨識方法。此方法使用臉部的RGB色彩資訊。由於使用臉部的RGB色彩資訊是屬於高維度的特徵空間，因此需藉由 median filer 和 downsampling 來減少特徵向量的維度; 再經過 singular value decomposition (SVD) 取出最後用來辨識的特徵向量。使用此特徵向量輸入至 support vector machine (SVM) 演算法中用以辨別此特徵向量是屬於真人人臉還是假人人臉。 我們的實驗結果與其他的傳統方法比較可以知道我們所提出的方法在使用較少的特徵下仍然擁有較高的辨識率。因此這個方法可以在較少的運算時間下達到高的辨識率，而且辨識率對於在臉部的表情變化及不同光源強度的變化下仍然優於其他方法。
然而，為了使真假人臉的辨識方法更能運作在未知的光源下，則獲得光源資訊可以改善真假人臉辨識是必須的且是重要的。 為了得知未知的光源資訊，是可以藉由低維度的線性模型和基底來描述光源，並且使用一般的 Gaussian-like 的CCD 感應器或者是 CMOS 感應器來求得光源模型的權重係數並用來重建光源。 我們的目標是從所有的感應器組合中找到最佳的一組感應器組合，並且用來重建光源。我們選擇的最佳的感應器組合是與最佳的 Gaussian 感應器有最好的關聯性。同時所選的最佳感應器組合所產生的矩陣必須有最小的 condition number 以減少計算誤差。
除此之外，利用反射係數不受光源改變及座標位置改變的特性，臉部的反射係數可以用來當作特徵向量，並將此特徵向量輸入至 Fisher linear discriminant (FLD) 用以作為真假人臉辨識之用。 而反射係數可以在已知的光源下利用三刺激值導出。此已知光源可以藉由上述的方法求得。 從結果可以知道這個用Fisher linear discriminant (FLD) 做為真假人臉辨識的方法可以比其他方法得到更好的辨識率。

This dissertation presents a novel algorithm for face discrimination between real and synthetic human faces. The facial features extracted from singular value decomposition (SVD) are very robust features that are invariant to the proportional variance of image intensity and rotation as well as have the characteristic of insensitiveness to image noise. The nonlinear support vector machine (SVM) is then used to determine whether the extracted features represent a real human face or a synthetic human face. Experimental results demonstrate that our method is shown to have a higher rate of success, and is less sensitive to facial expression variations and varying brightness levels than other methods.
Moreover, in order for face discrimination to work reliably under unusual lighting conditions, obtaining illuminant information can improve the performance of face discrimination. Therefore, we present a method of recovering spectra-daylight using these most proper color sensors chosen from color sensors of CCD camera or CMOS camera. The recovering spectra-daylight can achieve high spectral and colorimetric accuracy with a reduce number of spectral bands of color sensors.
Finally, the facial reflectance coefficients as additional facial feature vector are derived from the reflectance model based on illuminant information. These facial reflectance coefficients that have illumination and geometrical invariant properties are sent as an input vector to the Fisher linear discriminant (FLD) to solve the problem of discriminating between real and synthetic human faces. Some experimental results show the features extracted by our system can achieve high discrimination rate.

Reference

1. Chiunhisun Lin and Kuo-Chin Fan, “Triangle-based approach to the detection of human face”, Pattern Recognition, 34, 1271-1284 (2001).
2. Chin-Chuan Han, Hong-Yuan Mark Liao, and Gwo-Jong Yu, Liang-Hua Chen, “Fast face detection via morphology-based pre-processing,” Pattern Recognition, 33, 1701-1712 (2000).
3. Rein-Lien Hsu, Mohamed Abdel-Mottaleb, and Anil K. Jain, “Face detection in color image”, IEEE Trans. Pattern Analysis and Machine Intelligence, 24(5), 696-706 (May 2002).
4. Kin-Man Lam, and Hong Yan, “An analytic-to-holistic approach for face recognition”, IEEE Trans. Pattern Analysis and Machine Intelligence, 20(7), 673-785 (July 1998).
5. Li-Fen Chen, Hong-Yuan Mark Liao, Ja-Chen Lin, and Chin-Chuan Han, “Why recognition in a statistics-based face recognition system should be based on the pure face portion: probabilistic decision-based proof,” Pattern recognition, 34, 1393-1403, (2001).
6. Jian Yang, Jing-yu Yang, David Zhang, and Jian-feng Lu, “Feature fusion: parallel strategy vs. serial strategy,” Pattern recognition, 36, 1369-1381, (2003).
7. Hua Yu and Jie Yang, “A direct LDA algorithm for high-dimensional data – with
application to face discriminant,” Pattern recognition, 34, 2067-2070, (2001).
8. Huilin Xiong, M.N.S. Swamy, and M.O. Ahmad, “Two-dimensional FLD for face recognition,” Pattern recognition, 38, 1121-1124, (2005).
9. Jian Yang, Alejandro F. Frangi, and Jing-yu Yang, “A new kernel fisher discriminant algorithm with application to face recognition,” Neurocomputing, 56, 415-421, (2004).
10. Chengjun Liu and Harry Wechsler, “Gabor feature based classification using the enhanced fisher linear discriminant model for face recognition,” IEEE Trans. Image Processing, 11(4), 467-476, (April 2002).
11. Srinivas Gutta, Jeffrey R. J. Huang, P. Jonathon, and Harry Wechsler, “Mixture of experts for classification of gender, ethnic origin, and pose of human”, IEEE Trans. Neural Network, 11(4), 948-960 (July 2000).
12. Baback Moghaddam, and Ming-Hsuan Yang, “Learning gender with support faces,” IEEE Trans. Pattern Analysis and Machine Intelligence, 21(5), 707-711 (May 2002).
13. Kyunghee Lee, Yongwha Chung, and Hyeran Byun, “SVM-based face verification with feature set of small size”, Electronics Letter, 38(15), 787-789 (July 2002).
14. Robert E. Karlsen, David J. Gorsich, and Grant R. Gerhart, “Target classification via support vector machines,” Optical Engineering, 39(3), 704-711 (March 2000)
15. Olivier Chapelle, Patrick Haffiner, and Vladimir N. Vapnik, “Support vector machines for histogram-based image classification,” IEEE Trans. Neural Network, 10(5), 1055-1999 (September 1999).
16. Vojislav Kecman, “Learning and soft computing – support vector machines, neural networks, and fuzzy logic models,” A Bradford book (2001).
17. Valdimir N. Vapnik, “An overview of statistical learning theory,” IEEE Trans. Neural Network, 10(5), 988-999 (Sept 1999)
18. Vladimir N. Vapnik, “Statistical learning theory,” John Wiley & Sons, New York (1998).
19. Thorsten Joachims, “Learning to classify text using support vector machines,” Kluwer academic publishers (Feb 2001).
20. Yuan Tian, Tieniu Tan, Yunhong Wang, and Yuchun Fang, “Do singular values contain adequate information for face discriminant,” Pattern Recognition, 36, 649-655 (2003).
21. Quan Pan, Min-Gui Zhang, De-Long Zhou, Yong-Mei Cheng, and Hong-Cai Zhang, “Face recognition based on singular-value feature vectors,” Optical Engineering, 42(8), 2368-2374 (August 2003).
22. VIRGINIA C. Klema, “The singular value decomposition: its computation and some applications,” IEEE Trans. Autom. Control, AC-25, 164-176 (1980).
23. Zi-Quan Hong, “Algebraic feature extraction of image for recognition,” Pattern recognition, 24, 211-219 (1991).
24. Roger A. Horn, Charles R. Johnson, “Matrix analysis,” Cambridge University Press, Cambridge, 411-455 (1990).
25. Yong-Oing cheng, et al., “Human face discriminant method based on the statistical model of small sample size,” SPIE proceedings of the Intell. Robots and Comput. Vision, 1607, 85-95 (1991).
26. Yankiang Wang and Baozong Yuan, “A novel approach for human face detection from color images under complex background,” Pattern recognition, 34, 1983-1992, (2001).
27. Hongxun Yao and Wen Gao, “Face detection and location based on skin chrominance and lip chrominance transformation from color images,” Pattern recognition, 34, 1555-1564, (2001).
28. Zi-Quan Hong, “Algebraic feature extraction of image for recognition,” Pattern recognition, 24(3), 221-219, (1991).
29. Jian Huang Lai, Pong C. Yuen, Guo Can Feng, “Face recognition using holistic Fourier invariant features,” Pattern recognition, 34, 95-109, (2001).
30. John C. Platt, “Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods,” in Advances in Large Margin Classifiers, Cambridge, MA: MIT press, 1999.
31. William H. Press, Saul A. Teukolsky, William T. Vetterling, and Brian P. Flannery, “Numerical recipes in C: The art of scientific computing (2nd ed),” Cambridge University Press, Cambridge, 1992..
32. N.E. Ayat, M. Cheriet, C. Y. Suen, “Automatic model selection for the optimization of SVM kernels,” Pattern recognition, 38, 1733-1745, (2005).
33. David G. Luenberger, “Linear and Nonlinear Programming,” Addison-Wesley, Reading, 1984.
34. R. Fletcher, “Practical Methods of Optimization,” John Wiley and Sons,Chichester 1987.
35. Dimitri P. Bertsekas, “Nonlinear Programming,” Athena Scientific, Belmont,MA, 1995.
36. Jeffrey M. Dicarlo and Brian A. Wandell, “Illuminant estimation beyond the bases”, in Proceedings of Eihth Color Imaging Conference: Color Science, System, and Applications (Society for Imaging Science and Technology, Springfield, Va., 2000), 91-96.
37. Javier Hernandez-Andres, Juan L. Nieves, Eva M. Valero, and Javier Romero, “Spectral-daylight recovery by use of only a few sensors,” J. Opt. Soc. Am. A, Vol. 21, No. 1, 13-23, 2003.
38. DavidH. Marimont and BrianA. Wandell, “Linear models of surface and illuminant spectra,” J. Opt. Soc. Am. A, Vol. 9, 1905-1913, 1992.
39. Zhihong. Pan, Glenn. Healey, and David. Slater, “Global spectral irradiance variability and material discrimination at Boulder, Colorado,” J. Opt. Soc. Am. A, Vol. 20, No. 3, 513-521, 2003.
40. Javier Romero, Antonio Garcia-Beltran, and Javier Hernandez-Andres, “Linear bases for representation of natural and artificial illuminants,” J. Opt. Soc. Am. A, Vol. 14, No. 5, 1007-1014, 1997.
41. David S. Watkins, “Fundamentals of matrix computation,” John Wiley and Sons, Inc.,1991.
42. Shoji Tominaga, “Multichannel vision system for estimating surface and illumination functions”, J. Opt. Soc. Am. A, Vol. 13, No. 11, 2163-2173, 1996.
43. Shoji Tominaga, and Brian A. Wandell, “Natural scene-illuminant estimation using the sensor correlation,” Processdings of IEEE, Vol. 90, No. 1, 42-56, Jan 2002.
44. David Connah, Stephen Westland, and Mitchell G. A. Thomson, “Recovering spectral information using digital camera systems”, Color Technol. 117, 309–312 (2001).
45. David Connah, Stephen Westland, and Mitchell G. A. Thomson, “Optimization of a multispectral imaging system,” in Proceedings of the First European Conference on Colour in Graphics, Imaging and Vision (Society for Imaging Science and Technology, Springfield, Va., 2002), 619–622.
46. David Connah, Stephen Westland, and Mitchell. G. A. Thomson, “Parametric investigation of Multispectral imaging,” in 9th Congress of the International Color Association, R. Chung and A. Rodrigues, eds., Proc. SPIE 4421, 943–946 (2001).
47. J. J. Michalsky and E. W. Kleckner, ‘‘Estimation of continuous solar spectral distributions from discrete filter measurements,’’ Sol. Energy 33, 57–64 (1984).
48. J. J. Michalsky, ‘‘Estimation of continuous solar spectral distributions from discrete filter measurements: II. A demonstration of practicability,’’ Sol. Energy 34, 439–445 (1985).
49. Shuxue. Quan and Noboru Ohta, ‘‘Evaluating hypothetical spectral sensitivities with quality factors,’’ J. Imaging Sci. Technol. 46, 8–14 (2002).
50. Shuxue. Quan, Noboru Ohta, R. S. Berns, and X. Jiang, ‘‘Unified measure of goodness and optimal design of spectral sensitivity functions,’’ J. Imaging Sci. Technol. 46, 485–497 (2002).
51. J. Herna´ndez-Andre´s, J. Romero, J. L. Nieves, and R. L. Lee, Jr., ‘‘Color and spectral analysis of daylight in southern Europe,’’ J. Opt. Soc. Am. A 18, 1325–1335 (2001).
52. J. Herna´ndez-Andre´s, J. Romero, and R. L. Lee, Jr., ‘‘Colorimetric and spectroradiometric characteristics of narrowfield-of-view clear skylight in Granada, Spain,’’ J. Opt. Soc.Am. A 18, 412–420 (2001).
53. Juan L. Nieves, Eva M. Valero, Sergio M.C. Nascimento, Javier Hernandez-Andres, and Javier Romero, ‘‘Multispectral synthesis of daylight using a commercial digital CCD camera,’’ Applied Optics, Vol. 44, No. 27, 5696–5703 (2005)..
54. Michael J. Vrhel and H.Joel. Trussell, ‘‘Optimal color filters in the presence of noise,’’ IEEE Trans. Image Proessing Vol.4, No.6, 814–823 (1995).
55. Michael J. Vrhel and H. Joel Trussell, ‘‘Filter consideration in color correction,’’ IEEE Trans. Image Proessing Vol.3, No.2, 147–161 (1994).
56. Jan-Mark Geusebroek, Rein van den Boomgaard, Arnold W.M. Smeulders, and Hugo Geerts, “Color Invariance,” IEEE Trans. Pattern Analysis and Machine Intelligence, 23(12), 1338-1350 (Dec 2001).
57. John. Daugman, “An information-theoretic view of analog representation in striate cortex,” in Computational Neuroscience, E. Schwartz, Ed. Cambridge, MA: MIT Press, 1990, pp. 403–424.